fix: temporarily disable cuda-ipc for mm data tensor (sgl-project#8431)

Signed-off-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
Co-authored-by: Xinyuan Tong <xinyuantong.cs@gmail.com>

Author: 2 people

python/sglang/srt/multimodal/processors/base_processor.py

@@ -221,6 +221,13 @@ def process_mm_data(
             return_tensors="pt",
             **kwargs,
         )
+        # move feature tensors to cpu
+        for feature_name in self.FEATURE_NAMES:
+            if feature_name in result and isinstance(
+                result[feature_name], torch.Tensor
+            ):
+                result[feature_name] = result[feature_name].to("cpu")
+
         return result
 
     @abstractmethod
@@ -623,19 +630,4 @@ def process_and_combine_mm_data(
                 mm_token_id=mm_token_id,
             )
 
-        # post-process
-        for item in all_collected_items:
-            # replace the feature tensor with a proxy
-            if isinstance(item.feature, torch.Tensor) and item.feature.is_cuda:
-                item.feature = TransportProxyTensor(
-                    transport_mode=self.transport_mode, data=item.feature
-                )
-            elif (
-                isinstance(item.precomputed_embeddings, torch.Tensor)
-                and item.precomputed_embeddings.is_cuda
-            ):
-                item.precomputed_embeddings = TransportProxyTensor(
-                    transport_mode=self.transport_mode, data=item.precomputed_embeddings
-                )
-
         return all_collected_items, input_ids, ret